End of training

all_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0017900855411426165,
-    "train_runtime": 1355.0897,
-    "train_samples": 318,
-    "train_samples_per_second": 0.236,
-    "train_steps_per_second": 0.015
+    "train_loss": 0.0903062883833627,
+    "train_runtime": 775.6737,
+    "train_samples": 160,
+    "train_samples_per_second": 0.413,
+    "train_steps_per_second": 0.026
 }

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9d0fcf89b0d9fed5e9b396b90064dd869b504a22409166e53d8ebffdb5c43037
+oid sha256:712f3499dad1cde946b4845bd472421caa52bd2c96a9442ee63d6b82278c705e
 size 4996670464

model-00002-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e006d1cc424f76049ea432489e0a0571da7e79a9a2eb066c085800d68e1179f0
+oid sha256:74b14aea7e8a50e1dff59e3a0a55034489d5c1d6b5acae59f4f92b9378f826d9
 size 1178224960

train_results.json

@@ -1,8 +1,8 @@
 {
     "total_flos": 0.0,
-    "train_loss": 0.0017900855411426165,
-    "train_runtime": 1355.0897,
-    "train_samples": 318,
-    "train_samples_per_second": 0.236,
-    "train_steps_per_second": 0.015
+    "train_loss": 0.0903062883833627,
+    "train_runtime": 775.6737,
+    "train_samples": 160,
+    "train_samples_per_second": 0.413,
+    "train_steps_per_second": 0.026
 }

trainer_state.json

@@ -2,7 +2,7 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.5031446540880503,
+  "epoch": 1.0,
   "eval_steps": 500,
   "global_step": 20,
   "is_hyper_param_search": false,
@@ -10,203 +10,203 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "completion_length": 244.65625,
-      "epoch": 0.050314465408805034,
-      "grad_norm": 38.64108657836914,
+      "completion_length": 384.84375,
+      "epoch": 0.1,
+      "grad_norm": 45.04591751098633,
       "kl": 0.0,
       "learning_rate": 4.965903258506806e-07,
       "loss": -0.0,
-      "reward": 3.212035158649087,
-      "reward_std": 0.8912933750543743,
-      "rewards/concensus_correctness_reward_func": 0.32799999602138996,
-      "rewards/consensus_reward_func": 0.625,
+      "reward": 3.691977483453229,
+      "reward_std": 1.1356534652295522,
+      "rewards/concensus_correctness_reward_func": 0.6882499977946281,
+      "rewards/consensus_reward_func": 0.6875,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.625,
-      "rewards/question_recreation_reward_func": 0.6870976558420807,
+      "rewards/final_correctness_reward_func": 0.8125,
+      "rewards/question_recreation_reward_func": 0.6260087669361383,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.09375,
-      "rewards/xmlcount_reward_func": 0.8531875107437372,
+      "rewards/strict_format_reward_func": 0.15625,
+      "rewards/xmlcount_reward_func": 0.7214687555097044,
       "step": 2
     },
     {
-      "completion_length": 232.25,
-      "epoch": 0.10062893081761007,
-      "grad_norm": 17.972475051879883,
-      "kl": 0.07448420615401119,
+      "completion_length": 290.28125,
+      "epoch": 0.2,
+      "grad_norm": 56.04582595825195,
+      "kl": 0.048623996786773205,
       "learning_rate": 4.698684378016222e-07,
-      "loss": 0.0001,
-      "reward": 4.904678091406822,
-      "reward_std": 0.7286372808739543,
-      "rewards/concensus_correctness_reward_func": 0.5300000086426735,
-      "rewards/consensus_reward_func": 1.4375,
+      "loss": 0.0,
+      "reward": 5.547513723373413,
+      "reward_std": 0.9548060596571304,
+      "rewards/concensus_correctness_reward_func": 1.1509375181049109,
+      "rewards/consensus_reward_func": 1.0625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4375,
-      "rewards/question_recreation_reward_func": 0.9278343506157398,
+      "rewards/final_correctness_reward_func": 1.0,
+      "rewards/question_recreation_reward_func": 0.9249514192342758,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.328125,
-      "rewards/xmlcount_reward_func": 1.2437187507748604,
+      "rewards/strict_format_reward_func": 0.21875,
+      "rewards/xmlcount_reward_func": 1.1903750076889992,
       "step": 4
     },
     {
-      "completion_length": 283.9375,
-      "epoch": 0.1509433962264151,
-      "grad_norm": 36.8934326171875,
-      "kl": 0.7191745502641425,
+      "completion_length": 216.34375,
+      "epoch": 0.3,
+      "grad_norm": 23.09837532043457,
+      "kl": 892.624653175706,
       "learning_rate": 4.193203929064353e-07,
-      "loss": 0.0007,
-      "reward": 4.8152162581682205,
-      "reward_std": 1.0298546410776908,
-      "rewards/concensus_correctness_reward_func": 1.0431249924004078,
-      "rewards/consensus_reward_func": 1.125,
+      "loss": 0.8926,
+      "reward": 4.4300857884809375,
+      "reward_std": 1.2345747215877054,
+      "rewards/concensus_correctness_reward_func": 0.9410000052303076,
+      "rewards/consensus_reward_func": 0.75,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.625,
-      "rewards/question_recreation_reward_func": 0.6894663451239467,
+      "rewards/question_recreation_reward_func": 0.7900232495740056,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.203125,
-      "rewards/xmlcount_reward_func": 1.12950000166893,
+      "rewards/strict_format_reward_func": 0.25,
+      "rewards/xmlcount_reward_func": 1.074062503874302,
       "step": 6
     },
     {
-      "completion_length": 232.9375,
-      "epoch": 0.20125786163522014,
-      "grad_norm": 26.585630416870117,
-      "kl": 2.184791928390041,
+      "completion_length": 257.6875,
+      "epoch": 0.4,
+      "grad_norm": 25.814899444580078,
+      "kl": 1.4642285024747252,
       "learning_rate": 3.5042385616324236e-07,
-      "loss": 0.0022,
-      "reward": 4.871623918414116,
-      "reward_std": 1.4442174723371863,
-      "rewards/concensus_correctness_reward_func": 1.1043750010430813,
-      "rewards/consensus_reward_func": 0.875,
+      "loss": 0.0015,
+      "reward": 4.062486097216606,
+      "reward_std": 1.364403745283198,
+      "rewards/concensus_correctness_reward_func": 0.5777500020340085,
+      "rewards/consensus_reward_func": 0.625,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4375,
-      "rewards/question_recreation_reward_func": 0.9222801737487316,
+      "rewards/final_correctness_reward_func": 0.8125,
+      "rewards/question_recreation_reward_func": 0.7780798338353634,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.296875,
-      "rewards/xmlcount_reward_func": 1.2355937510728836,
+      "rewards/strict_format_reward_func": 0.171875,
+      "rewards/xmlcount_reward_func": 1.0972812492400408,
       "step": 8
     },
     {
-      "completion_length": 271.84375,
-      "epoch": 0.25157232704402516,
-      "grad_norm": 230.85166931152344,
-      "kl": 9.327558241784573,
+      "completion_length": 286.8125,
+      "epoch": 0.5,
+      "grad_norm": 17.246906280517578,
+      "kl": 2.0072783255018294,
       "learning_rate": 2.706448363680831e-07,
-      "loss": 0.0093,
-      "reward": 5.3308946043252945,
-      "reward_std": 1.1659031981835142,
-      "rewards/concensus_correctness_reward_func": 0.9190000081434846,
-      "rewards/consensus_reward_func": 0.8125,
+      "loss": 0.002,
+      "reward": 6.18870684504509,
+      "reward_std": 0.7351647131145,
+      "rewards/concensus_correctness_reward_func": 1.5492499843239784,
+      "rewards/consensus_reward_func": 1.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.3125,
-      "rewards/question_recreation_reward_func": 0.8955507688224316,
+      "rewards/final_correctness_reward_func": 1.25,
+      "rewards/question_recreation_reward_func": 0.929519459605217,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.234375,
-      "rewards/xmlcount_reward_func": 1.1569687500596046,
+      "rewards/strict_format_reward_func": 0.21875,
+      "rewards/xmlcount_reward_func": 1.2411874905228615,
       "step": 10
     },
     {
-      "completion_length": 205.125,
-      "epoch": 0.3018867924528302,
-      "grad_norm": 14.090639114379883,
-      "kl": 0.360865599475801,
+      "completion_length": 222.3125,
+      "epoch": 0.6,
+      "grad_norm": 38.915470123291016,
+      "kl": 3.321804000530392,
       "learning_rate": 1.886286282148002e-07,
-      "loss": 0.0004,
-      "reward": 4.681136727333069,
-      "reward_std": 0.8340875742869684,
-      "rewards/concensus_correctness_reward_func": 0.8726250045001507,
-      "rewards/consensus_reward_func": 0.8125,
+      "loss": 0.0033,
+      "reward": 4.649261876940727,
+      "reward_std": 1.167486259713769,
+      "rewards/concensus_correctness_reward_func": 1.1363750249147415,
+      "rewards/consensus_reward_func": 0.75,
       "rewards/cumulative_reward_2": 0.0,
       "rewards/final_correctness_reward_func": 0.625,
-      "rewards/question_recreation_reward_func": 0.9242929741740227,
+      "rewards/question_recreation_reward_func": 0.793293122202158,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.21875,
-      "rewards/xmlcount_reward_func": 1.2279687523841858,
+      "rewards/strict_format_reward_func": 0.28125,
+      "rewards/xmlcount_reward_func": 1.0633437521755695,
       "step": 12
     },
     {
-      "completion_length": 274.6875,
-      "epoch": 0.3522012578616352,
-      "grad_norm": 15.313384056091309,
-      "kl": 1.40120963845402,
+      "completion_length": 259.875,
+      "epoch": 0.7,
+      "grad_norm": 12.302684783935547,
+      "kl": 1.9860680536367,
       "learning_rate": 1.1326296046939333e-07,
-      "loss": 0.0014,
-      "reward": 6.080421447753906,
-      "reward_std": 0.8273013745201752,
-      "rewards/concensus_correctness_reward_func": 1.2539374958723783,
-      "rewards/consensus_reward_func": 1.375,
+      "loss": 0.002,
+      "reward": 4.9309286857023835,
+      "reward_std": 0.8483433704241179,
+      "rewards/concensus_correctness_reward_func": 1.1531875003129244,
+      "rewards/consensus_reward_func": 0.9375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 1.0625,
-      "rewards/question_recreation_reward_func": 0.958671472966671,
+      "rewards/final_correctness_reward_func": 0.8125,
+      "rewards/question_recreation_reward_func": 0.6953349150717258,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.203125,
-      "rewards/xmlcount_reward_func": 1.2271874994039536,
+      "rewards/strict_format_reward_func": 0.21875,
+      "rewards/xmlcount_reward_func": 1.113656248897314,
       "step": 14
     },
     {
-      "completion_length": 230.875,
-      "epoch": 0.4025157232704403,
-      "grad_norm": 42.175933837890625,
-      "kl": 1.0333388892468065,
+      "completion_length": 260.1875,
+      "epoch": 0.8,
+      "grad_norm": 12.547839164733887,
+      "kl": 0.4333856268785894,
       "learning_rate": 5.271487265090163e-08,
-      "loss": 0.001,
-      "reward": 4.291793182492256,
-      "reward_std": 0.8799512176192366,
-      "rewards/concensus_correctness_reward_func": 0.4721875060349703,
-      "rewards/consensus_reward_func": 1.25,
+      "loss": 0.0004,
+      "reward": 5.027402684092522,
+      "reward_std": 1.352033179719001,
+      "rewards/concensus_correctness_reward_func": 1.0091875093057752,
+      "rewards/consensus_reward_func": 0.9375,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.125,
-      "rewards/question_recreation_reward_func": 0.8352931793779135,
+      "rewards/final_correctness_reward_func": 0.75,
+      "rewards/question_recreation_reward_func": 0.8164025899022818,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.359375,
-      "rewards/xmlcount_reward_func": 1.249937504529953,
+      "rewards/strict_format_reward_func": 0.265625,
+      "rewards/xmlcount_reward_func": 1.2486874982714653,
       "step": 16
     },
     {
-      "completion_length": 232.59375,
-      "epoch": 0.4528301886792453,
-      "grad_norm": 19.220884323120117,
-      "kl": 0.406570911873132,
+      "completion_length": 258.15625,
+      "epoch": 0.9,
+      "grad_norm": 14.867289543151855,
+      "kl": 0.38399916142225266,
       "learning_rate": 1.3545689574841341e-08,
       "loss": 0.0004,
-      "reward": 4.516285641235299,
-      "reward_std": 0.5985990336630493,
-      "rewards/concensus_correctness_reward_func": 0.5337499966844916,
-      "rewards/consensus_reward_func": 1.0625,
+      "reward": 5.083541095256805,
+      "reward_std": 1.3432283122092485,
+      "rewards/concensus_correctness_reward_func": 1.1145625049248338,
+      "rewards/consensus_reward_func": 0.75,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.625,
-      "rewards/question_recreation_reward_func": 0.8423481268109754,
+      "rewards/final_correctness_reward_func": 1.0,
+      "rewards/question_recreation_reward_func": 0.7862598411738873,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.28125,
-      "rewards/xmlcount_reward_func": 1.1714375019073486,
+      "rewards/strict_format_reward_func": 0.234375,
+      "rewards/xmlcount_reward_func": 1.1983437463641167,
       "step": 18
     },
     {
-      "completion_length": 253.875,
-      "epoch": 0.5031446540880503,
-      "grad_norm": 21.807924270629883,
-      "kl": 2.3870232792105526,
+      "completion_length": 269.21875,
+      "epoch": 1.0,
+      "grad_norm": 76.03479766845703,
+      "kl": 0.8027685410343111,
       "learning_rate": 0.0,
-      "loss": 0.0024,
-      "reward": 5.31898732483387,
-      "reward_std": 1.3964124553604051,
-      "rewards/concensus_correctness_reward_func": 1.2262500124052167,
-      "rewards/consensus_reward_func": 1.4375,
+      "loss": 0.0008,
+      "reward": 4.308647858910263,
+      "reward_std": 0.8510435534408316,
+      "rewards/concensus_correctness_reward_func": 0.8827499886974692,
+      "rewards/consensus_reward_func": 1.0,
       "rewards/cumulative_reward_2": 0.0,
-      "rewards/final_correctness_reward_func": 0.4375,
-      "rewards/question_recreation_reward_func": 0.893924780189991,
+      "rewards/final_correctness_reward_func": 0.375,
+      "rewards/question_recreation_reward_func": 0.767147877253592,
       "rewards/soft_format_reward_func": 0.0,
-      "rewards/strict_format_reward_func": 0.125,
-      "rewards/xmlcount_reward_func": 1.1988125070929527,
+      "rewards/strict_format_reward_func": 0.203125,
+      "rewards/xmlcount_reward_func": 1.0806249976158142,
       "step": 20
     },
     {
-      "epoch": 0.5031446540880503,
+      "epoch": 1.0,
       "step": 20,
       "total_flos": 0.0,
-      "train_loss": 0.0017900855411426165,
-      "train_runtime": 1355.0897,
-      "train_samples_per_second": 0.236,
-      "train_steps_per_second": 0.015
+      "train_loss": 0.0903062883833627,
+      "train_runtime": 775.6737,
+      "train_samples_per_second": 0.413,
+      "train_steps_per_second": 0.026
     }
   ],
   "logging_steps": 2,